Measurement Systems Analysis

Transcription

1 Measurement Systems Analysis

2 Objectives At the end of the session, the students are expected to: Recognize how Minitab is used in measurement systems analysis Design a gage capability analysis for variables and attributes data

3 The importance of measurement systems How measurement systems affect data Variation in part + Variation in measurement = Total variation

4 Toolbox Type I Gage Study Repeatability and Bias Gage R&R Repeatability and Reproducibility Gage Linearity and Bias Study Linearity & Bias

5 Type I Gage Study Example Bear decides to buy a new digital vernier gauge. He wants to confirm how well his new vernier measures objects so he takes two metal blocks and measures their lengths on his CPU controlled laser profiling equipment. The first object has a length of 50 mm and Bear gives himself a tolerance width of 0.5 mm for measuring this block in the future. Bear measures the length of this object 50 times and records the results in Worksheet: Type 1GD. The results are in the Block 1 column. Use the results and carry out a Type 1 Gage Study to establish if the measurement system is repeatable and if a bias is present. Stat<<Quality Tools<<Gage Study<<Type 1 Gage Study

6 Type I Gage Study

7 Type I Gage Study

8 Type I Gage Study Exercise The second object has a length of 64.5mm and Bear gives himself a tolerance width of 0.8 mm for measuring this block in the future. Bear measures the length of this object 75 times and records the results in Worksheet: Type IGS. The results are in the Block 2 column. Use the results and carry out a Type 1 Gage study to establish if the measurement system is repeatable and if a bias is present. Is the measurement system suitable?

9 Gage Repeatability and Reproducibility Study Create a Randomized Study Worksheet for a GR&R Study Stat<<Quality Tools<<Gage study<<create Gage R&R Study Worksheet

10 Gage Repeatability and Reproducibility Study Create a Randomized Study Worksheet for a GR&R Study Automotive Industry Action Group (AIAG) recommend using: 10 parts 2 or 3 operators 2 or 3 replicates

11 Gage Repeatability and Reproducibility Study Classic Method: Metrics %Contribution %Study Variation %Tolerance %Process Number of Distinct Categories (NoDC) Assistant: Metrics %Tolerance %Process %Tolerance Tolerance: to improve parts to a specification %Study Variation or %Process Process: to reduce part-to-part variation

12 Automotive Industry Action Group %Contribution %Study Variation %Process %Tolerance Ideal <1% <10% Marginal/Acceptable 9% 30% Unacceptable >9% >30%

13 Gage R&R Study Example My three kids, Iqra, Humzah and Raeesa are not very good at baking cakes. I give them 10 packets of flour ask them to weight the packets as a part of a GR&R Study. Each of my kids measures each packet three times, so there are 10 parts, three operators, three repeats of each measurement making 90 runs in total. My kids have short attention spans so I did not need to randomize the study. The cake baking process at my house does not have a specification so I choose not to use a tolerance for this analysis. Worksheet: Cakes

14 Gage R&R Study Example Stat<<Quality Tools<< Gage Study<<Gage Run Chart

15 Gage R&R Study Example

16 Gage R&R Study Example Stat<<Quality Tools<< Gage Study<<Gage R&R Study (Crossed)

17 Gage R&R Study Example

18 Gage R&R Study Example Part-Person Interaction : dependent upon the part the person s ability to measure is

19 Gage R&R Study Example

20 Gage R&R Study Example Bards has set up a study examining the measurement system used when measuring the caliper of plasterboard. He uses 3 operators and 10 parts. The parts represent the variation seen within the process and not the extremes of the process. The tolerance for the caliper is 0.6 mm. This MSA is part of a project which is intended to reduce process waste when the product goes out of the tolerance required by the customer. Bards has also collected historical data on these measurements to allow a more precise calculation of partto-part variation. The study was originally randomized but has been ordered for our convenience. Stat<<QT<<GS<<Gage Run Chart

21 Gage R&R Study

22 Gage R&R Study Get the standard deviation of historical data (for part-to to-part variation) Gage R&R Study (Crossed)<<Options

23 Gage R&R Study

24 Gage R&R Study

25 Gage R&R Study Exercise Omer is running an improvement project where he wants to improve a metric against specification levels. He sets up a study examining the measurement system used when measuring the density of fuel pallets. He uses 3 operators and 10 parts. The parts represent the variation seen within the process. The tolerance for the product is 0.6 kg/m 3. The historical standard deviation is given as Check the results using a Gage Run Chart. Then using the results given in Worksheet: Fuel Density, carry out a GR&R to establish if the measurement system is deemed to be acceptable for tolerance. Marginal is not good enough within the nuclear industry.

26 Gage R&R Study (Nested) Example Mahwish has set up a study examining the measurement system used when measuring the fire resistance of plasterboard. The measurement given is time taken for the plasterboard to fail when exposed to a set flame. She uses three operators and six parts, each part is split into three pieces. The same operator tests the pieces from the same part. The parts represent typical parts within the process. Worksheet: Fire Test

27 Gage R&R Study (Nested)

28 Gage R&R Study (Nested) Exercise Han has set up a study examining the measurement system used when measuring the impact resistance of steel. He uses 3 operators and 9 parts, and each part is split into three pieces. The same operator tests the pieces from the same part. The parts represent the variation seen within the process. The tolerance for this measurement system is 3kN and the aim of this study is to improve parts to the tolerance. Worksheet: Impact Test

29 Gage Bias and Linearity Study Assumptions At least 5 parts that cover the operating range of the gage must be used. There must be at least 10 readings for each part The parts must be reference parts with known readings for each part It is best if the runs can be randomized and the parts kept blind

30 Gage Bias and Linearity Study Example Uzzy wants to carry out a Gage Bias and Linearity Study on the measurement system used to measure the caliper of plasterboard. The reference parts representing each of the standard sizes of board are used. The reference parts are 7.5, 9.5, 12.5, 15.0 and 19.0 mm. Each one is measured 12 times in a randomized study. It is known that six times the historical standard deviation is Worksheet: Caliper B&L Stat<<QT<<GS<<Gage Linearity and Bias Study

31 Gage Bias and Linearity Study

32 Gage Bias and Linearity Study Exercise Parky wants to carry out a Gage bias and linearity study on a weigh scale. Reference weights are used which cover the range of the weigh scale. These are 100, 200, 300, 400 and 500 kg. Each one is measured 10 times in a randomized study. Process variation is 1.5 kg. Using the results, carry out a Gage bias and linearity study to learn if the measurement system is suitable. Worksheet: Weights B&L

33 Attribute Agreement Analysis Example: A bank uses manual underwriting to analyze mortgage loan applications. Each underwriter must use the information supplied by the applicant and other external information such as credit history to classify the application into one of foue categories: decline or or do not fund, fund-1, fund-2, and fund-3. The fund-2 and fund-3 categories are applicants that are considered low-risk loans while fund-1 is a higher risk applicant. Suppose that 30 applications are selected and evaluated by a panel of senior underwriters who arrive at a consensus evaluation for each application, then three different underwriters (Sue, Fred and John) are asked to evaluate each application twice. The applications are blinded (customer names, addresses and other identifying information removed) and the two evaluations are performed several days apart. Refer to Worksheet: Attrib